Show overview
Seventy3 has been publishing since 2024, and across the 2 years since has built a catalogue of 737 episodes. That works out to roughly 220 hours of audio in total. Releases follow a near-daily cadence.
Episodes typically run ten to twenty minutes — most land between 14 min and 21 min — though episode length varies meaningfully from one episode to the next. None of the episodes are flagged explicit by the publisher. It is catalogued as a ZH-CN-language Technology show.
The show is actively publishing — the most recent episode landed yesterday, with 274 episodes already out so far this year. The busiest year was 2025, with 366 episodes published. Published by 任雨山.
Episodes
737
Running
2024–2026 · 2y
Median length
17 min
Cadence
Near-daily
From the publisher
73播客，名字取材于Sheldon最喜欢的数字，内容由NotebookLM生成，每天跟随AI读AI业界论文。
Latest Episodes
View all 737 episodes【第731期】AgentThread：大模型代理协议的形式化安全分析与验证
Sep 30, 202616 min
【第730期】InterSAGE：智能体互联网的安全性与可验证互操作协议
Sep 29, 202628 min
【第729期】Agent文件系统代理存储：组织、演化与可持续性研究
Sep 28, 202615 min
【第728期】ACM：面向长时任务的自主语境管理框架
Sep 27, 202624 min
【第727期】JAXBench：自主TPU内核优化评测基准
Sep 26, 202625 min
【第726期】Molt：面向代理强化学习的可扩展原生PyTorch训练框架
Sep 25, 202625 min
【第725期】隐性推理：大模型中不可见的计算过程
Sep 24, 202613 min
【第724期】ReOPD：多轮交互的重播前缀在线策略蒸馏
Sep 23, 202626 min
【第723期】NOOA：基于面向对象编程的原生Python智能体框架
Sep 22, 202621 min
【第722期】RoboTTT：通过测试时训练扩展机器人策略上下文
Sep 21, 202616 min
【第721期】结构化输出引发的语言模型答案同质化研究
Sep 20, 202617 min
【第720期】智能代理的长文本进阶披露机制评估
Sep 19, 202625 min
【第719期】GAMUT：大语言模型长文本生成事实完备性评价基准
Sep 18, 202617 min
【第718期】大语言模型中的全球工作空间：可言语化表征研究
Sep 17, 202623 min
【第717期】PRO-LONG：程序化记忆助力长程推理
Sep 16, 202612 min
【第716期】从记忆到技能：长程大语言模型智能体的循证协同演化治理
Sep 15, 202622 min
【第715期】Harness Handbook：AI智能体框架行为化表征指南
Sep 14, 202615 min
【粉丝投稿001期】HydroGym：流体力学强化学习通用平台
Sep 14, 202620 min
【第714期】ProxyMark：基于代理陷阱与流量水印的门罗币Tor节点去匿名化研究
Sep 13, 202623 min
【第713期】LingBot-World-Infinity：无限交互式现实世界模拟器
Sep 12, 202624 min
任雨山@小宇宙App